v 2010 IBM 1 Topas and nmon Nigel
v © 2010 IBM 1 Topas and nmon Nigel Griffiths Advanced Technical Support IBM Europe
v © 2009 IBM s pa to nmon for Linux now Open Source § nmon for Linux on Power then added x 86, Z, x 86_64 § Released by IBM - July 2009 – http: //nmon. sourceforge. net § Compile for your Linux flavour – Cell blades, disk engines, – Diskless, embedded ++ § On a Linux Distro's !! 2
v © 2009 IBM … nmon function within topas –AIX 5. 3 TL 09 - TL 11, + –AIX 6. 1 TL 02 - TL 04, + October 2008 s pa to nmon for AIX 4 … add AIX 5 & AIX 6 … nmon 12 e for AIX 3
v s pa to topas problem 1 © 2009 IBM 4 30+ pages, no product, no marketing, Redbook nope
v KB-W 5 AQW MWT PSz Vcsw %bdon MPuse hpit physb pmem s pa to topas problem 2 © 2009 IBM RTFM
v to 10 different online modes s pa topas problem 3 © 2009 IBM 7 file formats (with friends) 6
v © 2009 IBM s pa to Modes Topas Local AIX mode Cross-partition Mode [CEC/whole machine] Default View Detailed View 2 1 WPAR VIOS View 7 3 5 6 AMS 9 nmon Save 10 8 nmon Classic for AIX&Linux nmon View LPAR View CEC Save nmon mode WPAR nmon Save CSV file Other Local Save Tools topasout Text Reports 7 nmon Analyser nmon Consolidator nmon 2 rrd/nmon 2 web Other tools or databases
v © 2009 IBM s pa to Local AIX Topas VIOS view 8
v © 2009 IBM s pa to topas VIOS 2. 1+ monitoring Hit “E” for shared Ethernet adapter(s) SEA Direction of traffic in this case from Virtual Ethernet out of the VIOS § ent 2 virtual 14 KB/s incoming § ent 0 = physical = 14 KB/s outgoing § ent 10 = SEA = 14 KB/s in and 14 KB/s out Topas Monitor for host: bronze_ivm Interval: 2 Fri Dec 5 08: 19 2008 ======================================== Network KBPS I-Pack O-Pack KB-In KB-Out ent 10 (SEA) 31. 3 43. 0 15. 9 15. 4 |--ent 2 (VETH) 15. 9 15. 5 27. 0 14. 3 1. 6 --ent 0 (PHYS) 15. 5 27. 5 16. 0 1. 6 13. 9 lo 0 0. 0 VOIS [SEA] Physical Network 13. 9 KB/s 31. 3 KB/s Client Virtual Network 14. 3 KB/s 9
v © 2009 IBM § § § § § s pa to Another Example 10 Topas Monitor for host: bluevio 44 Interval: 2 Mon Nov 9 18: 28: 12 2009 ======================================== Network KBPS I-Pack O-Pack KB-In KB-Out ent 5 (SEA) 30. 9 35. 5 35. 0 15. 5 |--ent 1 (PHYS) 15. 5 25. 0 10. 5 14. 7 0. 8 --ent 4 (VETH) 15. 5 10. 5 24. 5 0. 8 14. 6 ent 6 (SEA) 1. 9 15. 5 16. 0 0. 9 1. 0 |--ent 0 (PHYS) 1. 0 15. 5 0. 9 0. 1 --ent 3 (VETH) 0. 9 0. 0 15. 5 0. 0 0. 9 lo 0 0. 0
v © 2009 IBM s pa to topas VIOS 2. 1+ monitoring 11 Hit “D” and then “d” You can see VIOS virtual/physical Disks And which hosts adapters is generating the I/O Topas Adapter View : bronze_ivm Interval: 2 Fri Dec 5 08: 29: 17 2008 ======================================== Adapter KBPS TPS KB-R KB-W sissas 0 9. 9 K 92. 0 0. 0 9. 9 K vhost 0 0. 0 vhost 1 9. 9 K 184. 0 92. 0 ======================================== Vtargets/Disks Busy% KBPS TPS KB-R ART MRT KB-W AWT MWT AQW AQD hdisk 0 0. 0 3. 6 0. 0 62. 6 0. 0 hdisk 1 25. 0 9. 9 K 92. 0 0. 0 3. 5 9. 9 K 4. 2 28. 9 0. 0 hdisk 2 0. 0 0. 0 hdisk 3 0. 0 0. 0 hdisk 4 0. 0 0. 0 Virtual Disks Client vhost Server vhost 1 Logical Volume sissas 0 hdisk 0 to 5
v © 2009 IBM s pa to Local AIX Workload Partition (WPAR) views 12
PA 13 R topas WLM active © 2008 13 IBM v W IBM System p U at pd e Topas Monitor for host: Fri Feb 20 08: 09: 25 2009 CPU ALL User% 48. 1 Network Total Kern% 1. 0 Wait% 0. 0 gold_lpar 3 Interval: 2 Idle% 50. 9 Entc 49. 7 KB-In 0. 6 KB-Out 0. 5 KBPS 1. 0 I-Pack 6. 0 Busy% 0. 0 KBPS 0. 0 TPS KB-Read KB-Writ 0. 0 KBPS 18. 1 TPS KB-Read KB-Writ 5. 0 18. 1 0. 0 Disk Total File. System Total WLM-Class (Active) System Default Name ncpu topas java wlmsched O-Pack 4. 5 Physc 0. 25 CPU% 1 0 Mem% 16 7 PID CPU% Pg. Sp Class 561208 47. 8 0. 1 wp 16 454778 0. 2 2. 0 System 241848 0. 2 48. 7 System 69666 0. 1 0. 5 System Blk-I/O% 0 0 EVENTS/QUEUES Cswitch 144 Syscall 156 Reads 6 Writes 1 Forks 0 Execs 0 Runqueue 1. 0 Waitqueue 0. 0 PAGING Faults Steals Pgsp. In Pgsp. Out Page. In Page. Out Sios 0 0 0 0 NFS (calls/sec) Ser. V 2 0 Cli. V 2 0 Ser. V 3 0 Cli. V 3 0 FILE/TTY Readch Writech Rawin Ttyout Igets Namei Dirblk MEMORY Real, MB % Comp % Noncomp % Client 18521 113 0 4 0 2048 38. 8 7. 7 PAGING SPACE Size, MB 512 % Used 1. 1 % Free 99. 9 WPAR Activ 3 WPAR Total 9 Press: "h"-help "q"-quit Note: WLM Section show up if on WPAR starts WLM automatically © 2009 IBM Corporation Workload Partition & Workload Partition Manager IBM Systems
v 14 U at pd topas then @ WPAR mode © 2008 14 IBM R -- or -- PA topas -@ W IBM System p e Topas Monitor for host: Fri Feb 20 08: 12: 01 2009 CPU ALL User% 48. 0 Network Total Kern% 1. 3 Wait% 0. 0 gold_lpar 3 Interval: 2 Idle% 50. 7 Entc 49. 9 KB-In 2. 5 KB-Out 2. 6 KBPS 5. 1 I-Pack 18. 5 Busy% 0. 0 KBPS 0. 0 TPS KB-Read KB-Writ 0. 0 KBPS 36. 2 TPS KB-Read KB-Writ 10. 0 36. 2 0. 0 Disk Total File. System Total WPAR wp 16 wp 17 Name ncpu java topas wlmsched O-Pack 18. 5 Physc 0. 25 CPU% 19 0 PID CPU% Pg. Sp 561208 47. 7 0. 1 241848 0. 3 48. 8 454778 0. 2 2. 0 69666 0. 1 0. 5 Mem% 1 1 Blk-I/O% 0 0 Wpar wp 16 Global EVENTS/QUEUES Cswitch 196 Syscall 206 Reads 14 Writes 1 Forks 0 Execs 0 Runqueue 1. 5 Waitqueue 0. 0 PAGING Faults Steals Pgsp. In Pgsp. Out Page. In Page. Out Sios 0 0 0 0 NFS (calls/sec) Ser. V 2 0 Cli. V 2 0 Ser. V 3 0 Cli. V 3 9 FILE/TTY Readch Writech Rawin Ttyout Igets Namei Dirblk MEMORY Real, MB % Comp % Noncomp % Client 37042 143 0 9 0 2048 38. 8 7. 7 PAGING SPACE Size, MB 512 % Used 1. 1 % Free 99. 9 WPAR Activ 3 WPAR Total 9 Press: "h"-help "q"-quit Conclusions: not really helpful – you many have a dozen WPARs Workload Partition & Workload Partition Manager © 2009 IBM Corporation IBM Systems
PA 15 R topas inside WPAR (W@ not possible) © 2008 15 IBM v W IBM System p U at pd e Topas Monitor for host: Fri Feb 20 08: 34: 25 2009 wp 16 Interval: 2 Kernel User Wait Idle Physc = 0. 6 99. 4 0. 0 0. 30 |# | |##############| | | %Resc= 59. 6 Network Total KBPS 0. 1 I-Pack 0. 5 File. System Total Name ncpu topas sendmail cron rpc. lock init srcmstr syslogd O-Pack 0. 5 KBPS 0. 0 PID CPU% 31981576 99. 4 34603052 0. 6 44564492 0. 0 5242886 0. 0 66060306 0. 0 1 0. 0 27262980 0. 0 43515918 0. 0 KB-In 0. 0 KB-Out 0. 1 TPS KB-Read KB-Writ 0. 0 Pg. Sp 0. 1 1. 0 1. 1 0. 5 0. 3 0. 6 0. 8 0. 3 EVENTS/QUEUES Cswitch 18 Syscall 37 Reads 0 Writes 0 Forks 0 Execs 0 Runqueue 0. 0 Waitqueue 0. 0 PAGING Faults Steals Pgsp. In Pgsp. Out Page. In Page. Out Sios 1741 0 0 0 NFS (calls/sec) Ser. V 2 0 Cli. V 2 0 Ser. V 3 0 Cli. V 3 0 FILE/TTY Readch Writech Rawin Ttyout Igets Namei Dirblk MEMORY Real, MB % Comp % Noncomp % Client 0 89 0 337 0 0 0 2048 37. 7 PAGING SPACE Size, MB 512 % Used 1. 1 % Free 99. 9 Press: "h"-help "q"-quit Undocumented! Conclusions: WPAR can’t see other processes of Partition the CPU used by© 2009 the. IBMWPAR ? ? Corporation Workload Partition 99% & Workload Manager IBM Systems
© 2008 16 IBM 16 R PA WPAR detailed mode topas -W@ -- or -topas + @W v W IBM System p U at pd e Topas Monitor for host: WPAR wp 16 wp 18 wp 17 gold_lpar 3 CPU% 19 0 0 Interval: Mem% 1 1 1 2 Fri Feb 20 08: 25: 11 2009 Blk-I/O% 0 0 0 19%= of a CPU Entitlement=0. 5 47. 7% = of entitlement ======================================= DATA TEXT PAGE PGFAULTS USER PID PRI NI RES SPACE TIME CPU% I/O OTH COMMAND root 561260 688216 64 0 34 3 34 0: 00 47. 7 0 0 ncpu root 688216 520340 60 20 146 68 146 0: 00 0. 3 1 85 ksh root 680122 532560 58 41 445 102 445 0: 00 0. 2 0 0 topas root 241848 225458 71 20 12496 22 12496 0: 13 0. 2 0 0 java root 69666 0 16 41 128 0: 00 0. 1 0 0 wlmsched root 446544 1 60 20 464 0: 00 0. 1 0 0 nfsd root 65568 0 37 41 240 0: 00 0. 0 0 0 gil pconsole 274590 327848 71 20 13072 18 13072 0: 13 0. 0 0 0 java root 323744 123068 60 20 646 263 646 0: 00 0. 0 0 0 rmcd root 540922 471156 60 20 273 184 273 0: 00 0. 0 0 0 sendmail Conclusions: improving Workload Partition & Workload Partition Manager © 2009 IBM Corporation IBM Systems
v © 2008 17 IBM 17 R PA U nmon classic for AIX 612 Global W IBM System p at pd e ┌─nmon 12 f─────#=PURR Stats───────Host=gold_lpar 3─────Refresh=2 secs───09: 16. 43─┐ │ CPU-Utilisation-Small-View ──────Entitled. CPU= 0. 50 Used. CPU= 0. 347─────│ │Logical CPUs 0 -----25 ------50 -----75 -----100│ │CPU User% Sys% Wait% Idle%| | |│ │ 0 29. 7 4. 3 0. 5 65. 6|UUUUUUUss >│ │ 1 0. 0 100. 0| > |│ │ 2 0. 0 100. 0|> |│ │ 3 0. 0 100. 0|> |│ │Entitle. Capacity/Virtual. CPU +-----------|------|------+│ │ EC 60. 9 7. 9 0. 0 0. 6|UUUUUUUUUUUUUUUsss--------|│ │ VP 15. 2 2. 0 0. 2|UUUUUUU---------------------|│ │EC= 69. 4% VP= 17. 4% +--No Cap---|--Folded=1 --|------100% VP=2 CPU+│ │ Workload Partition (WPAR) ──────────────────────────│ │Number of WPARs= 9 | Physical-CPU|Memory-in-MB | │ │WPAR-Name Type State | User System|In. Use numperm| Run. Q p. Switch Fork │ │wp 17 Sys Active | 0. 00| 17. 8 3. 9| 0. 0 0 0. 0 │ │wp 18 Sys Active | 0. 00| 17. 3 3. 9| 0. 0 0 0. 0 │ │wp 15 Sys Defined | | | │ │wp 16 Sys Active | 0. 29 0. 00| 21. 4 6. 2| 1. 0 13 0. 0 │ │wp 21 Sys Defined | | | │ │wp 22 Sys Defined | | | │ │wp 23 Sys Defined | | | │ │wp 20 Sys Defined | | | │ │wp 07 Sys Defined | | | │ │ Work-Load-Manager (WLM) ───────────────────────────│ │Classes=3 Mode=Active Rsets Proc. Total Class. Total │ │ CPU MEM BIO │ │Class Name |---Used-Percent--||--Desired-||----Shares-----|Proc's T I L │ │wp 16 22. 5 1. 2 0. 0 100 98 100 -1 -1 -1 19 0 0 0 │ │wp 17 0. 0 1. 0 0. 0 100 98 100 -1 -1 -1 16 0 0 0 │ │wp 18 0. 0 1. 0 0. 0 100 98 100 -1 -1 -1 16 0 0 0 │ │Total percentage 22. 5 3. 3 0. 0 Tier, Inheritance & Local. Shm=T I L │ │───────────────────────────────────────│ © 2009 IBM Corporation Workload Partition & Workload Partition Manager IBM Systems Conclusions: Understandable
v © 2009 IBM s pa to CEC Whole machine Cross Partition All LPAR view 18
v © 2009 IBM s pa to Cross Partition View HMC § LPAR: topas –C § VIOS Over network Once only Can take 20 seconds – topas then hit “C” – topas –cecview § Totally different program inside topas xmtopas xmwlm xmperf § Only LPARs on same sub-net – Edit $HOME/Rsi. hosts – Fully hostname. domain one per line Over network Once every 10 seconds Linux AIX AIX X 19
v © 2009 IBM s pa to Active Memory Sharing (AMS) view 20
v © 2009 IBM s pa to Monitoring 1 LPAR for AMS is insane 21
v s pa AMS Have to monitor all LPARs to topas –C © 2009 IBM No details for Linux on POWER or IBM i = no daemon Regular extra “g” details Topas CEC Monitor Partition Info Monitored : 4 Un. Monitored: Shared : 4 Uncapped : 4 Capped : 0 Dedicated : 0 Donating : 0 Interval: 10 Memory (GB) Processor Monitored : 8. 0 Monitored : 2. 0 Un. Monitored: Available : Un. Allocated: Consumed : 6. 5 Shared : 2 Dedicated : 0 Donated : 0 Pool Size : 4 Wed Dec 3 10: 15: 06 2008 Virtual Pools : 0 Avail Pool Proc: 3. 7 Shr Physical Busy: 0. 28 Ded Physical Busy: 0. 00 Donated Phys. CPUs 0. 00 Stolen Phys. CPUs : 0. 00 Hypervisor Virt. Context Switch: 976 Phantom Interrupts : 1 Host OS M Mem In. U Lp Us Sy Wa Id Phys. B Vcsw Ent %Ent. C Ph. I pmem -------------------shared------------------silver_vios 1 A 61 U 2. 0 1. 9 8 08 1 0 90 0. 01 873 0. 40 2. 6 0 silver_lpar 2 A 61 UM 2. 0 1. 4 2 49 1 0 49 0. 26 244 0. 50 51. 3 1 1. 30 silver_lpar 3 A 61 UM 2. 0 1. 6 2 0 0 0 99 0. 01 294 0. 50 1. 4 0 0. 78 silver_lpar 4 A 61 UM 2. 0 1. 7 2 0 0 0 99 0. 01 220 0. 50 1. 3 0 0. 87 silver_lpar 5 A 61 UM 2. 0 1. 7 2 0 0 0 99 0. 01 218 0. 50 1. 1 0 1. 01 Inu In. Use LPAR Working Set Logical Memory Mem Dedicated Memory Shared memory LPAR Physical Memory pmem 22
v © 2009 IBM Topas CEC Monitor Interval: 10 Partitions Memory (GB) Memory Pool(GB) Mshr: 4 Mon: 8. 0 In. Use: 6. 8 Mpsz: 4. 0 MPuse: 4. 0 Mded: 0 Avl: 1. 2 Pools: 1 s pa to AMS Pool Level: topas –C then hit “m” Wed Dec 3 10: 56: 12 2008 I/O Memory(GB) Entl: 308. 0 Use: 47. 9 mpid mpsz mpus memu iome iomu hpit ----------------------------------------0 4. 00 3. 99 8. 00 6. 82 308. 0 47. 9 32 31 Hypervisor Page-in Time (msecs) Hypervisor Page-in/second Pool Size Pool Used Machine memory Used In use I/O Memory assign for device driver DMA I/O For more on I/O Memory see: • lparstat -me 23
v © 2009 IBM s pa to Mode Changing 24 ~ topas default DETAIL C a Nigel’s Monitor nmon topas_nmon mode ~ CEC topas -C
§ Freely available § Performance monitor for – AIX 5 and 6 – AIX 4 (via older nmon version) – Linux – POWER, x 86 & mainframe § Design principles – Zero installation time & simple to use – For the performance expert – max info on the screen – Less than 1% CPU – For large machines – 64 CPUs, 4000+ disks, 35, 000 processes v nmon basics © 2010 IBM 25
v nmon – Online or Data Collector © 2010 IBM 26 nmon CSV file in nmon format Online Data Collection
nmon Stephen’s nmon file v Screen -f or –F option © 2010 IBM nmon 27 Consolidator Stephen’s Others Federico’s p. Graph Bruce’s nmon 2 web Perl nmon 2 rrd C filter Java Dynamic Graphs nmon Analyser Excel graphs Scripts rrdtool scripts + CPU & RAM aggregation - Create rrd - Load rrd - Graph rrd index. html nmon flow Website +. jpg graphs rrdtool open source
v © 2010 IBM 28 nmon is half the story The other half is due to the nmon Analyser & nmon Consolidator Excel spreadsheets from Stephen Atkins, UK another personal time project
v © 2010 IBM 29 nmon Consolidator v 1. 4. 1 § Produces overview charts for CPU, Memory, Network, Disk – Fast & simple code, pre-generated charts § Multiple nmon/topasout files for: – ALL LPARs of one machine – Clustered system – Oracle RAC / HPC – Modelling changes from Dedicated to Shared CPU LPARs – Server consolidation Modelling – Reporting actual or potential savings from virtualisation § Trend charts - processing multiple files from 1 LPAR
v © 2010 IBM 30 AIX Labs in 2008: “nmon will be assimilated!”
v © 2010 IBM 31 Code stuffing Lines of code from 6500 9000
v nmon on POWER 6 & AIX 6 + New Features for V 12 User Requested Advanced, POWER 6 and AIX 6 items 1. Disk Service Times 8. Virtual I/O Server SEA via entstat 2. Selecting Particular Disks 9. Partition Mobility 3. Time Drift 10. WPAR & Application Mobility 4. Multiple Page Sizes 5. 4 KB, 64 KB, ++ 11. Dedicated Donating (POWER 6) Timestamps in UTC & no. of digits 6. More Kernel & Hypervisor Stats * 12. Folded CPU count 7. High Priority nmon 13. Multiple Shared Pools (POWER 6) © 2010 IBM 32 (POWER 6) (AIX 6) (SPLPAR) Housekeeping items 14. Fibre Channel stats via fcstat 15. Bug fixes – small or fine tuning, see nmon wiki for a list 16. Network packet sizes now saved to file 17. Warnings from nmon like: network overflow 18. TOP Online - sizes in KB, MB, GB to keep columns aligned 19. Fast abort in capture mode – SIGUSR 1 during config collection stops nmon 20. Return an error & on invalid options – return codes to check it finished normally 21. Adapter "not available" work around – see nmon FAQ wiki about these AIX bugs 22. EMC hdiskpower renamed to “power” to save screen space 23. NFS v 4 for AIX 5. 3 ML 5+
v © 2010 IBM 33 Direct Fibre-Channel and VIOS SEA Shared Ethernet with ‘O’ Fibre Channel with ‘^’ also regular LPAR & FC tape drives § Actually using more expensive entstat and fcstat commands ┌─nmon 12──────h=Help───────Host=silver_vios 1───Refresh=2 secs───10: 03. 36─┐ │ │ │ Shared-Ethernet-Adapter ───────────────────────────│ │ Adapter Receive Transmit -Requests--Size KB-│ │Number Name KB/s In Out │ │ 1 ent 3 141. 4 211. 6 22. 5 67. 0 1024. 0 │ │ Totals 0. 0 MB/s 21. 6 22. 5 │ │ Fibre-Channel-Adapter ────────────────────────────│ │ Adapter Receive Transmit -Requests--Size KB-│ │Number Name KB/s In Out │ │ 1 fcs 0 9945. 8 0. 0 420. 2 0. 0 17. 9 │ │ Totals 0. 0 MB/s 0. 0 │ │───────────────────────────────────────│ │ │
v © 2010 IBM 34 nmon topas_nmon
v © 2010 IBM 35 topas_nmon Install Clarification: Do nothing as it is Installed by Default with AIX
v © 2010 IBM 36 topas nmon Command Typical place users but the nmon shell script $ nmon script /usr/bin/nmon binary /usr/bin/topas_nmon hard link to topas & topasrec $ topas and hit ~ $ nmon –f –s 300 –c 288 Just works as normal
v Using: smitty topas © 2010 IBM 37 # smitty topas ----------------------Add Host to topas external subnet search file ( Rsi. hosts) List hosts in topas external subnet search file ( Rsi. hosts) List active Recordings Start Persistent local recording Start New Recording Start Persistent CEC recording Stop Recording Start local recording List completed recordings Start CEC recording Generate Report Period Day Hour Custom Start nmon recording [Entry Fields] Type of Recording nmon * Length of Recording custom * Recording Interval in seconds * Number of Samples * Overwrite existing recording Output Path Running Priority * Recording time zone Include top processes Include … … Binary nmon Start now & once only Via Cron every day Start Persistent XXXXX recording Type or select values in entry fields. Press Enter AFTER making all desired changes. [300] # [288] # no + [] [] # Local time+ top process+ * * [Entry Fields] Type of Recording Length of Recording persistent Recording Interval in seconds Number of Days to store per file Number of Days to retain Output Path Overwrite existing recording file CEC [60] # [1] # [31] # [] no +
topas nmon Classic nmon v © 2010 IBM 38
v © 2010 IBM 39 nmon Support – AIX 5. 3 TL 09+ & AIX 6. 1 TL 02+ – You have full AIX Support with SWMA – Raise a PMR topas nmon Use new Official nmon – Google: AIX “Performance Tool Forum” nmon – Google: AIX nmon wiki Classic Older AIX releases use nmon Classic – Please ONLY version 12 e+
nmon classic – development stopped ALL new development in topas_nmon topas nmon topas nmon topas nmon Classic nmon Development 12 e topas nmon v © 2010 IBM 40 12 f
v © 2010 IBM 41 nmon is now GREEN fully recycled source code No animals harmed in the testing of nmon
v © 2010 IBM 42 And Finally the End 1) Topas with nmon mode worth learning 2) Google: nmon wiki 3) Running nmon Classic use the latest version 4) Uncapped Shared CPU LPARs Monitor Physical CPU use Note Utilisation as 100% = Entitlement
v Blatant Advert for Hands-On Movies AIX 6 & Power 6 - the movie! 10 on WPAR 8 on AIX 7 on POWER 6 - SPLPAR/SMT 10 on IVM 5 on Cool stuff nmon 15 on Power. VM 3 on AMS 12 on System Director Each movie 5 to 40 minutes Downloadable files in. wmv format Watch actual use of these new features In total ~15 hours AIX movie © 2010 IBM 43
v Ganglia = “nmon” Cross Partitions Whole machine: Physical CPU used © 2010 IBM 44 Last Hour Last Day Last Week Last Month Open Source collections: CPU, Disks, Memory, Network, Run Queue + POWER stats: Entitlement, VP, weights Physical CPU use Enhanced Code: http: //www. perzl. org/ganglia/ Docs: http: //www. ibm. com/developerworks/wikis/display/Wiki. Ptype/ganglia Last Quarter
Default View v Topas Demo Notes © 2010 IBM 45 1) c n d f p – cool!!! = CPU Network Disk Filesystem Processes 2) Left over is called the Fixed Section 3) P for Processes it is now a larger list CEC or Cross Partitions View or Whole Machine 4) c CPUs c graph 1) topas –C or topas and hit C c ALL 2) s and d Shared CPU and Dedicated CPU sections c Off 3) g Global 5) n networks N Totals 4) m Memory pool = AMS stats from Hypervisor N Off (select 0 hit f – note: only 1 pool) 6) f filesystems f Totals 5) p CPU pool stats (I only have one) f Off 6) v VIO Server/Client disk use f to select the VIOS 7) d Disks D Totals D Off 8) If present t = tape, w=WLM and @=WPARs 9) a=Reset All Detailed View 1) D Disks in Full Detail D disks d a. Dapters (if VIOS virtual adapters) d a. Dapters v virtual reorder columns Exit back to D Down to a adapter and hit f m Multi-path I/O for the disks attached 2) E Ethernet 3) F Filesystems 4) L LPAR and individual CPUs e AMD LPAR stats (could show this on silver_lpar 4) 5) P Processes Page space – don’t have this else where Show moving columns with and 6) T Tape if there is a ATAPE device attached 7) V Volume group (seems to have a problem on VIOS 2. 1 – restart topas) Down to a vg then hit f LV in the VG 8) W WLM then @ WPAR Down to a WPAR and hit f Also in Process, Filesystem Detailed views
- Slides: 45